DS595Project4Mario

To train a model, simply follow the example in train.py. The reward types are "default", "wrapper", and "sparse".

Model weight files

The model weight files are too big to include in this repository. For the full repo with the models, please download at https://drive.google.com/file/d/14qNBTIYYh7gpnagyCgUZyVXO77mfp-jw/view?usp=sharing

Name		Name	Last commit message	Last commit date
Latest commit History 33 Commits
default_DDQN		default_DDQN
default_DQN		default_DQN
default_PER_DDQN		default_PER_DDQN
default_PER_DQN		default_PER_DQN
figures		figures
figures_longer		figures_longer
sparse_DDQN		sparse_DDQN
sparse_DQN		sparse_DQN
sparse_PER_DDQN		sparse_PER_DDQN
sparse_PER_DQN		sparse_PER_DQN
wrapper_DQN		wrapper_DQN
wrapper_PER_DDQN		wrapper_PER_DDQN
wrapper_PER_DQN		wrapper_PER_DQN
.gitignore		.gitignore
Experiment_sheet.xlsx		Experiment_sheet.xlsx
PER.py		PER.py
README.md		README.md
a3c_model.py		a3c_model.py
ddqn_agent.py		ddqn_agent.py
ddqn_agent_PER.py		ddqn_agent_PER.py
dqn_agent.py		dqn_agent.py
dqn_agent_PER.py		dqn_agent_PER.py
dqn_agent_y.py		dqn_agent_y.py
dqn_model.py		dqn_model.py
dqn_model_y.py		dqn_model_y.py
mario_wrapper.py		mario_wrapper.py
mellowmax_agent.py		mellowmax_agent.py
mellowmax_agent_PER.py		mellowmax_agent_PER.py
mm_default.png		mm_default.png
mm_per_default.png		mm_per_default.png
mm_per_sparse.png		mm_per_sparse.png
mm_per_wrapper.png		mm_per_wrapper.png
mm_sparse.png		mm_sparse.png
mm_wrapper.png		mm_wrapper.png
test.py		test.py
testM.py		testM.py
train.py		train.py
train_thread.py		train_thread.py